1 Processamento dos dados

source("./load_files.R")
data_files_selected <- 
  list.files(
    pattern = ".*thais_.*csv$",
    recursive = TRUE,
    ignore.case = TRUE
    )

data_files_and_size <- 
  sapply(data_files_selected, file.size)

files_to_include_in_dataframe <- 
  tibble(
    "Files" = names(data_files_and_size),
    "Size (in MB)" = data_files_and_size/1E6
   )

skimr::skim(files_to_include_in_dataframe)
Data summary
Name files_to_include_in_dataf…
Number of rows 24
Number of columns 2
_______________________
Column type frequency:
character 1
numeric 1
________________________
Group variables None

Variable type: character

skim_variable n_missing complete_rate min max empty n_unique whitespace
Files 0 1 128 130 0 24 0

Variable type: numeric

skim_variable n_missing complete_rate mean sd p0 p25 p50 p75 p100 hist
Size (in MB) 0 1 2.18 3.48 0.32 0.38 0.6 2.1 10.99 ▇▁▁▁▁ 
cdr <- load_cdr(names(data_files_and_size))

cdr %<>% 
  mutate(
    expgroup = case_when(
      str_detect(file, "thaisnovoheader") ~ "nh",
      str_detect(file, "thais_29") ~ "29",
      str_detect(file, "thais_66") ~ "66",
      TRUE ~ "unknown"),
    cycle = case_when(
      str_detect(file, "R0_R2") ~ "R0_R2",
      str_detect(file, "R0_R3") ~ "R0_R3",
      str_detect(file, "R0_R4") ~ "R0_R4",
      str_detect(file, "R2_R3") ~ "R2_R3",
      str_detect(file, "R2_R4") ~ "R2_R4",
      str_detect(file, "R3_R4") ~ "R3_R4",
      TRUE ~ "unknown"),
    time = case_when(
        str_detect(file, "Initial") ~ "initial",
        str_detect(file, "Final") ~ "final",
        TRUE ~ "unknown")) %>% 
  select(cdr3, cycle, time, expgroup, everything())

cdr %<>% 
  group_by(cdr3, expgroup, cycle) %>% 
  arrange(desc(time), .by_group = TRUE) %>% 
  mutate(
    fcp = cdrp / lag(cdrp, default = first(cdrp)),
    fcq = quantity / lag(quantity, default = first(quantity))
  ) %>% 
  select(cdr3:quantity, fcp, fcq, everything())
cdr %>% 
  filter(time == "final") %>% 
  # filter(str_detect(cycle, "R0")) %>% 
  group_by(expgroup, cycle, time) %>% 
  arrange(desc(fcp)) %>% 
  slice_head(prop = .1) %>% 
  # slice_head(n = 1000) %>% 
  ggplot(aes(expgroup, log10(fcp))) +
    geom_violin(aes(fill = expgroup, color = expgroup), alpha = 0.5) +
    geom_jitter(aes(shape = expgroup), alpha = 0.6, size = 1) +
    stat_summary(
      fun = mean,
      fun.min = mean,
      fun.max = mean,
      geom = "crossbar",
      # width = 0.5,
      aes(color = expgroup)
    ) +
    facet_grid(. ~ cycle)

# cdr %>% 
#   filter(str_detect(cycle, "R0")) %>% 
#   filter(time == "final") %>% 
#   group_by(expgroup, cycle, time) %>% 
#   arrange(desc(fcp)) %>% 
#   slice_head(n = 1000) %>% 
#   ggplot(aes(fcp, color = expgroup, fill = expgroup)) +
#     geom_density(stat = "bin", alpha = 0.3) +
#     facet_grid(cycle ~ expgroup)

cdr %<>% 
  group_by(expgroup, cycle, time) %>% 
  arrange(desc(fcp)) %>% 
  slice_head(prop = .1) %>% 
  mutate(
    threshold = mean(log10(fcp))
    ) %>% 
  mutate(
    rich = if_else(
        (log10(fcp) >= threshold) &
          # (time == "final") &
          # (str_detect(cycle, "R0")),
          (time == "final"),
      "rich",
      "medium")) %>% 
  full_join(cdr) %>% 
  mutate(
    rich = if_else(
      is.na(rich),
      "poor",
      rich)) %>% 
  mutate(
    rich = factor(rich,
                  levels = c("rich", "medium", "poor"))
    ) %>% 
  mutate(
    threshold = if_else(
      is.na(threshold),
      0,
      threshold
    )
  )

cdr %>% 
  filter(rich == "rich") %>%
  ggplot() +
    geom_violin(aes(expgroup, log10(fcp), fill = expgroup)) +
    geom_jitter(aes(expgroup, log10(fcp), shape = expgroup), alpha = .2) +
    facet_grid(rich ~ cycle)

cdr %>% 
  filter(!rich == "rich") %>%
  ggplot() +
    geom_violin(aes(expgroup, log10(fcp), fill = expgroup)) +
    geom_jitter(aes(expgroup, log10(fcp), shape = expgroup), alpha = .2) +
    facet_grid(rich ~ cycle)

names(cdr)
##  [1] "cdr3"      "cycle"     "time"      "expgroup"  "cdrp"      "quantity" 
##  [7] "fcp"       "fcq"       "length"    "MW"        "AV"        "IP"       
## [13] "flex"      "gravy"     "SSF_Helix" "SSF_Turn"  "SSF_Sheet" "n_A"      
## [19] "n_C"       "n_D"       "n_E"       "n_F"       "n_G"       "n_H"      
## [25] "n_I"       "n_K"       "n_L"       "n_M"       "n_N"       "n_P"      
## [31] "n_Q"       "n_R"       "n_S"       "n_T"       "n_V"       "n_W"      
## [37] "n_Y"       "aliphatic" "aromatic"  "neutral"   "positive"  "negative" 
## [43] "invalid"   "file"      "threshold" "rich"
cdr %>% 
  filter(rich == "rich") %>% 
  filter(expgroup == "29" & cycle == "R0_R4") %>% 
  ungroup() %>% 
  select(cdr3:SSF_Sheet & !c("cycle", "time", "expgroup"))
## # A tibble: 13 x 14
##    cdr3     cdrp quantity    fcp    fcq length    MW    AV    IP  flex   gravy
##    <chr>   <dbl>    <int>  <dbl>  <dbl>  <int> <dbl> <dbl> <dbl> <dbl>   <dbl>
##  1 GPPH… 1.65e-1    23551 1.42e4 2.36e4     13 1590. 0.231  5.97 0.735 -1.26  
##  2 DLHN… 1.32e-1    18816 1.14e4 1.88e4      9 1132. 0.222  4.20 0.739 -0.967 
##  3 RYRN… 5.64e-2     8067 4.87e3 8.07e3     19 2410. 0.368  5.96 0.744 -1.71  
##  4 AYPE… 1.47e-1    21041 2.12e3 3.51e3     10 1212. 0.3    4.05 0.714 -0.2   
##  5 ITTV… 3.66e-3      524 3.16e2 5.24e2     14 1691. 0.286  5.83 0.702 -0.0857
##  6 RGSS… 1.66e-3      237 1.43e2 2.37e2      9 1005. 0.222  5.84 0.782 -1.12  
##  7 DY    3.36e-4       48 5.80e0 9.60e0      2  296. 0.5    4.05 0.728 -2.4   
##  8 ETWG… 5.59e-5        8 4.83e0 8.00e0      7  881. 0.286  4.05 0.774 -1.7   
##  9 DV    3.22e-4       46 2.78e0 4.60e0      2  232. 0      4.05 0.746  0.35  
## 10 QQIA… 2.80e-5        4 2.41e0 4.00e0      7  850. 0.143  4.05 0.723 -0.243 
## 11 GTSS… 2.10e-5        3 1.81e0 3.00e0     11 1213. 0.182  4.05 0.742 -0.191 
## 12 GVMG… 2.10e-5        3 1.81e0 3.00e0     10 1215. 0.4    4.05 0.678  0     
## 13 GYSS… 2.10e-5        3 1.81e0 3.00e0      9 1049. 0.333  4.05 0.759 -1.43  
## # … with 3 more variables: SSF_Helix <dbl>, SSF_Turn <dbl>, SSF_Sheet <dbl>
cdr %>% 
  filter(rich == "poor") %>% 
  filter(expgroup == "66" & cycle == "R0_R4") %>% 
  ungroup() %>% 
  select(cdr3:SSF_Sheet & !c("cycle", "time", "expgroup")) %>% 
  ggplot() +
  geom_density(aes(MW))

cdr %>% 
  filter(rich == "rich") %>% 
  filter(expgroup == "66" & cycle == "R0_R4") %>% 
  ungroup() %>% 
  select(cdr3:SSF_Sheet & !c("cycle", "time", "expgroup")) -> rich66

rich66
## # A tibble: 363 x 14
##    cdr3     cdrp quantity   fcp   fcq length    MW     AV    IP  flex   gravy
##    <chr>   <dbl>    <int> <dbl> <dbl>  <int> <dbl>  <dbl> <dbl> <dbl>   <dbl>
##  1 SQDV… 0.0138      2794  860.  1397     13 1440. 0.0769  4.05 0.733  0.0769
##  2 TAGV… 0.00433      876  539.   876      8  886. 0.25    4.05 0.718 -0.075 
##  3 GRPE… 0.00294      595  366.   595     10 1128. 0.1     5.32 0.764 -1.06  
##  4 GDLV… 0.00240      486  299.   486     13 1588. 0.308   4.21 0.732 -0.477 
##  5 ARPK… 0.00233      471  290.   471     14 1461. 0.0714  6.00 0.794 -1.26  
##  6 GRWI… 0.00206      417  257.   417      8  992. 0.25    8.75 0.711 -0.338 
##  7 DLDW… 0.00175      355  218.   355      9 1199. 0.333   4.05 0.705  0.244 
##  8 SGAV… 0.00173      351  216.   351     13 1414. 0.154   4.37 0.721  0.331 
##  9 TRKQ… 0.00163      330  203.   330      8  957. 0      11.0  0.801 -1.84  
## 10 DGRY… 0.00153      310  191.   310      8 1063. 0.375   4.21 0.741 -1.9   
## # … with 353 more rows, and 3 more variables: SSF_Helix <dbl>, SSF_Turn <dbl>,
## #   SSF_Sheet <dbl>
names(cdr)
##  [1] "cdr3"      "cycle"     "time"      "expgroup"  "cdrp"      "quantity" 
##  [7] "fcp"       "fcq"       "length"    "MW"        "AV"        "IP"       
## [13] "flex"      "gravy"     "SSF_Helix" "SSF_Turn"  "SSF_Sheet" "n_A"      
## [19] "n_C"       "n_D"       "n_E"       "n_F"       "n_G"       "n_H"      
## [25] "n_I"       "n_K"       "n_L"       "n_M"       "n_N"       "n_P"      
## [31] "n_Q"       "n_R"       "n_S"       "n_T"       "n_V"       "n_W"      
## [37] "n_Y"       "aliphatic" "aromatic"  "neutral"   "positive"  "negative" 
## [43] "invalid"   "file"      "threshold" "rich"
cdr %>% 
  filter(expgroup == "66" & cycle == "R0_R4") %>% 
  ggplot() +
  geom_violin(aes(rich, gravy, color = rich)) +
  geom_jitter(aes(rich, gravy), alpha = .1)

2 Tentativa de Clusterização

2.1 Clusterization with Quantity and Prevalence variables

2.1.1 Clusterização Experimento 29

library(Rtsne)
set.seed(42)

tsne_df <- cdr %>% 
  # filter(!rich == "poor") %>%
  filter(time == "final") %>% 
  filter(cycle == "R0_R4") %>% 
  filter(expgroup == "29")

tsne_df %>% 
  summarise(across(everything(), ~ all(sum(is.na(.x))))) %>% 
  select(where(isTRUE))
## # A tibble: 1 x 2
## # Groups:   expgroup, cycle [1]
##   expgroup cycle
##   <chr>    <chr>
## 1 29       R0_R4
tsne_df %>% 
  filter(is.na(threshold)) %>% 
  select(rich, threshold, quantity)
## # A tibble: 0 x 6
## # Groups:   expgroup, cycle, time [0]
## # … with 6 variables: expgroup <chr>, cycle <chr>, time <chr>, rich <fct>,
## #   threshold <dbl>, quantity <int>
set.seed(42)
tsne_out <- 
  tsne_df %>% 
    ungroup() %>% 
    select(!where(is.character)) %>% 
    select(!c(which(apply(., 2, var)==0))) %>% 
    unique() %>% 
    Rtsne(
      X = .,
      dims = 3,
      perplexity = 420,
      theta = 0.1,
      max_iter = 2E3,
      verbose = T,
      pca_center = T,
      pca_scale = T,
      normalize = T,
      eta = 200.0,
      exaggeration_factor = 12.0,
      num_threads = parallel::detectCores() - 2
      )
## Performing PCA
## Read the 1544 x 42 data matrix successfully!
## OpenMP is working. 6 threads.
## Using no_dims = 3, perplexity = 420.000000, and theta = 0.100000
## Computing input similarities...
## Building tree...
## Done in 4.72 seconds (sparsity = 0.932734)!
## Learning embedding...
## Iteration 50: error is 44.145871 (50 iterations in 7.22 seconds)
## Iteration 100: error is 44.145871 (50 iterations in 8.36 seconds)
## Iteration 150: error is 44.145871 (50 iterations in 9.83 seconds)
## Iteration 200: error is 44.145871 (50 iterations in 11.66 seconds)
## Iteration 250: error is 44.145871 (50 iterations in 14.65 seconds)
## Iteration 300: error is 1.193916 (50 iterations in 12.47 seconds)
## Iteration 350: error is 0.319184 (50 iterations in 6.44 seconds)
## Iteration 400: error is 0.241505 (50 iterations in 5.90 seconds)
## Iteration 450: error is 0.229975 (50 iterations in 6.36 seconds)
## Iteration 500: error is 0.226926 (50 iterations in 6.31 seconds)
## Iteration 550: error is 0.226181 (50 iterations in 6.22 seconds)
## Iteration 600: error is 0.225912 (50 iterations in 6.29 seconds)
## Iteration 650: error is 0.225838 (50 iterations in 6.83 seconds)
## Iteration 700: error is 0.225819 (50 iterations in 6.45 seconds)
## Iteration 750: error is 0.225798 (50 iterations in 6.39 seconds)
## Iteration 800: error is 0.225781 (50 iterations in 6.56 seconds)
## Iteration 850: error is 0.225778 (50 iterations in 6.94 seconds)
## Iteration 900: error is 0.225771 (50 iterations in 6.69 seconds)
## Iteration 950: error is 0.225767 (50 iterations in 6.80 seconds)
## Iteration 1000: error is 0.225760 (50 iterations in 6.82 seconds)
## Iteration 1050: error is 0.225743 (50 iterations in 6.83 seconds)
## Iteration 1100: error is 0.225722 (50 iterations in 6.36 seconds)
## Iteration 1150: error is 0.225723 (50 iterations in 6.43 seconds)
## Iteration 1200: error is 0.225719 (50 iterations in 6.22 seconds)
## Iteration 1250: error is 0.225719 (50 iterations in 6.67 seconds)
## Iteration 1300: error is 0.225725 (50 iterations in 6.38 seconds)
## Iteration 1350: error is 0.225728 (50 iterations in 6.39 seconds)
## Iteration 1400: error is 0.225729 (50 iterations in 6.37 seconds)
## Iteration 1450: error is 0.225728 (50 iterations in 6.50 seconds)
## Iteration 1500: error is 0.225730 (50 iterations in 6.48 seconds)
## Iteration 1550: error is 0.225729 (50 iterations in 6.37 seconds)
## Iteration 1600: error is 0.225730 (50 iterations in 6.39 seconds)
## Iteration 1650: error is 0.225734 (50 iterations in 6.81 seconds)
## Iteration 1700: error is 0.225733 (50 iterations in 6.59 seconds)
## Iteration 1750: error is 0.225732 (50 iterations in 6.42 seconds)
## Iteration 1800: error is 0.225733 (50 iterations in 6.30 seconds)
## Iteration 1850: error is 0.225730 (50 iterations in 6.48 seconds)
## Iteration 1900: error is 0.225729 (50 iterations in 6.46 seconds)
## Iteration 1950: error is 0.225730 (50 iterations in 6.30 seconds)
## Iteration 2000: error is 0.225735 (50 iterations in 6.35 seconds)
## Fitting performed in 284.28 seconds.
tsne_df %>% 
  ungroup() %>% 
  select(!where(is.character)) %>% 
  select(!c(which(apply(., 2, var)==0))) %>% 
  unique() -> a

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  ggplot() +
    geom_point(aes(V1, V2, color = a$rich))

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  plot_ly(
    title = "Sample title",
    x = .$V1,
    y = .$V2,
    z = .$V3,
    type = "scatter3d",
    mode = "markers",
    color = a$rich
  ) %>% 
  layout(title = "Experiment 29")
tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  plot_ly(
    title = "Sample title",
    x = .$V1,
    y = .$V2,
    z = .$V3,
    type = "scatter3d",
    mode = "markers",
    color = a$rich
  ) %>% 
  layout(title = "Experiment 29")

2.1.2 Clusterização Experimento 66

library(Rtsne)
set.seed(42)

tsne_df <- cdr %>% 
  # filter(!rich == "poor") %>%
  filter(time == "final") %>% 
  filter(cycle == "R0_R4") %>% 
  filter(expgroup == "66")

tsne_df %>% 
  summarise(across(everything(), ~ all(sum(is.na(.x))))) %>% 
  select(where(isTRUE))
## # A tibble: 1 x 2
## # Groups:   expgroup, cycle [1]
##   expgroup cycle
##   <chr>    <chr>
## 1 66       R0_R4
tsne_df %>% 
  filter(is.na(threshold)) %>% 
  select(rich, threshold, quantity)
## # A tibble: 0 x 6
## # Groups:   expgroup, cycle, time [0]
## # … with 6 variables: expgroup <chr>, cycle <chr>, time <chr>, rich <fct>,
## #   threshold <dbl>, quantity <int>
set.seed(42)
tsne_out <- 
  tsne_df %>% 
    ungroup() %>% 
    select(!where(is.character)) %>% 
    select(!c(which(apply(., 2, var)==0))) %>% 
    unique() %>% 
    Rtsne(
      X = .,
      dims = 3,
      perplexity = 30,
      theta = 0.5,
      max_iter = 1E3,
      verbose = T,
      pca_center = T,
      pca_scale = T,
      normalize = T,
      eta = 200.0,
      exaggeration_factor = 12.0,
      num_threads = parallel::detectCores() - 2
      )
## Performing PCA
## Read the 8436 x 42 data matrix successfully!
## OpenMP is working. 6 threads.
## Using no_dims = 3, perplexity = 30.000000, and theta = 0.500000
## Computing input similarities...
## Building tree...
## Done in 7.15 seconds (sparsity = 0.015493)!
## Learning embedding...
## Iteration 50: error is 95.444063 (50 iterations in 18.25 seconds)
## Iteration 100: error is 92.761432 (50 iterations in 24.37 seconds)
## Iteration 150: error is 91.017719 (50 iterations in 12.25 seconds)
## Iteration 200: error is 90.890341 (50 iterations in 10.68 seconds)
## Iteration 250: error is 90.831025 (50 iterations in 10.93 seconds)
## Iteration 300: error is 2.973973 (50 iterations in 8.21 seconds)
## Iteration 350: error is 2.540478 (50 iterations in 7.78 seconds)
## Iteration 400: error is 2.310407 (50 iterations in 8.07 seconds)
## Iteration 450: error is 2.166549 (50 iterations in 7.52 seconds)
## Iteration 500: error is 2.067423 (50 iterations in 7.54 seconds)
## Iteration 550: error is 1.995633 (50 iterations in 7.60 seconds)
## Iteration 600: error is 1.942256 (50 iterations in 8.12 seconds)
## Iteration 650: error is 1.902282 (50 iterations in 8.24 seconds)
## Iteration 700: error is 1.871328 (50 iterations in 8.12 seconds)
## Iteration 750: error is 1.847140 (50 iterations in 8.41 seconds)
## Iteration 800: error is 1.828978 (50 iterations in 8.32 seconds)
## Iteration 850: error is 1.815785 (50 iterations in 8.41 seconds)
## Iteration 900: error is 1.805099 (50 iterations in 8.40 seconds)
## Iteration 950: error is 1.796396 (50 iterations in 8.66 seconds)
## Iteration 1000: error is 1.788672 (50 iterations in 8.51 seconds)
## Fitting performed in 198.40 seconds.
tsne_df %>% 
  ungroup() %>% 
  select(!where(is.character)) %>% 
  select(!c(which(apply(., 2, var)==0))) %>% 
  unique() -> a

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  ggplot() +
    geom_point(aes(V1, V2, color = a$rich))

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  plot_ly(
    x = .$V1,
    y = .$V2,
    z = .$V3,
    type = "scatter3d",
    mode = "markers",
    color = a$rich
  ) %>% 
  layout(title = "Experiment 66")

2.2 Clusterization only with Biochemistry Properties

2.2.1 Experiment 29

library(Rtsne)
set.seed(42)

tsne_df <- cdr %>% 
  # filter(!rich == "poor") %>%
  filter(time == "final") %>% 
  filter(cycle == "R0_R4") %>% 
  filter(expgroup == "29")

tsne_df %>% 
  summarise(across(everything(), ~ all(sum(is.na(.x))))) %>% 
  select(where(isTRUE))
## # A tibble: 1 x 2
## # Groups:   expgroup, cycle [1]
##   expgroup cycle
##   <chr>    <chr>
## 1 29       R0_R4
tsne_df %>% 
  filter(is.na(threshold)) %>% 
  select(rich, threshold, quantity)
## # A tibble: 0 x 6
## # Groups:   expgroup, cycle, time [0]
## # … with 6 variables: expgroup <chr>, cycle <chr>, time <chr>, rich <fct>,
## #   threshold <dbl>, quantity <int>
set.seed(42)
tsne_out <- 
  tsne_df %>% 
    ungroup() %>% 
    select(cdr3, !where(is.character)) %>% 
    select(!c(which(apply(., 2, var)==0))) %>% 
    select(!c("quantity", "cdrp", "fcp", "fcq", "rich", "threshold")) %>% 
    unique() %>% 
    mutate(cdr3 = as.factor(cdr3)) %>% 
    Rtsne(
      X = .,
      dims = 3,
      perplexity = 30,
      theta = 0.5,
      max_iter = 2E3,
      verbose = T,
      pca_center = T,
      pca_scale = T,
      normalize = T,
      partial_pca = T,
      eta = 200.0,
      exaggeration_factor = 12.0,
      num_threads = parallel::detectCores() - 2
      )
## Performing PCA
## Read the 1589 x 50 data matrix successfully!
## OpenMP is working. 6 threads.
## Using no_dims = 3, perplexity = 30.000000, and theta = 0.500000
## Computing input similarities...
## Building tree...
## Done in 0.48 seconds (sparsity = 0.079423)!
## Learning embedding...
## Iteration 50: error is 73.611075 (50 iterations in 2.26 seconds)
## Iteration 100: error is 69.273657 (50 iterations in 1.37 seconds)
## Iteration 150: error is 68.914566 (50 iterations in 1.44 seconds)
## Iteration 200: error is 68.813606 (50 iterations in 1.50 seconds)
## Iteration 250: error is 68.765041 (50 iterations in 1.61 seconds)
## Iteration 300: error is 1.384633 (50 iterations in 1.31 seconds)
## Iteration 350: error is 1.190552 (50 iterations in 1.25 seconds)
## Iteration 400: error is 1.123723 (50 iterations in 1.29 seconds)
## Iteration 450: error is 1.099565 (50 iterations in 1.31 seconds)
## Iteration 500: error is 1.083687 (50 iterations in 1.37 seconds)
## Iteration 550: error is 1.075679 (50 iterations in 1.38 seconds)
## Iteration 600: error is 1.066637 (50 iterations in 1.38 seconds)
## Iteration 650: error is 1.056727 (50 iterations in 1.37 seconds)
## Iteration 700: error is 1.049040 (50 iterations in 1.44 seconds)
## Iteration 750: error is 1.042673 (50 iterations in 1.52 seconds)
## Iteration 800: error is 1.037077 (50 iterations in 1.27 seconds)
## Iteration 850: error is 1.033224 (50 iterations in 1.29 seconds)
## Iteration 900: error is 1.029217 (50 iterations in 1.27 seconds)
## Iteration 950: error is 1.026656 (50 iterations in 1.30 seconds)
## Iteration 1000: error is 1.024118 (50 iterations in 1.53 seconds)
## Iteration 1050: error is 1.021757 (50 iterations in 1.41 seconds)
## Iteration 1100: error is 1.020240 (50 iterations in 1.43 seconds)
## Iteration 1150: error is 1.018609 (50 iterations in 1.38 seconds)
## Iteration 1200: error is 1.016400 (50 iterations in 1.38 seconds)
## Iteration 1250: error is 1.015295 (50 iterations in 1.38 seconds)
## Iteration 1300: error is 1.013907 (50 iterations in 1.39 seconds)
## Iteration 1350: error is 1.012659 (50 iterations in 1.37 seconds)
## Iteration 1400: error is 1.011199 (50 iterations in 1.39 seconds)
## Iteration 1450: error is 1.009526 (50 iterations in 1.39 seconds)
## Iteration 1500: error is 1.008536 (50 iterations in 1.41 seconds)
## Iteration 1550: error is 1.007399 (50 iterations in 1.42 seconds)
## Iteration 1600: error is 1.005997 (50 iterations in 1.43 seconds)
## Iteration 1650: error is 1.004614 (50 iterations in 1.42 seconds)
## Iteration 1700: error is 1.003587 (50 iterations in 1.66 seconds)
## Iteration 1750: error is 1.002543 (50 iterations in 1.35 seconds)
## Iteration 1800: error is 1.001491 (50 iterations in 1.31 seconds)
## Iteration 1850: error is 1.000320 (50 iterations in 1.36 seconds)
## Iteration 1900: error is 0.999214 (50 iterations in 1.43 seconds)
## Iteration 1950: error is 0.998373 (50 iterations in 1.29 seconds)
## Iteration 2000: error is 0.997141 (50 iterations in 1.29 seconds)
## Fitting performed in 56.36 seconds.
tsne_df %>% 
  ungroup() %>% 
  select(cdr3, !where(is.character)) %>% 
  select(!c(which(apply(., 2, var)==0))) %>% 
  select(!c("quantity", "cdrp", "fcp", "fcq", "threshold")) %>% 
  unique() -> a

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  ggplot() +
    geom_point(aes(V1, V2, color = a$rich))

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  plot_ly(
    x = .$V1,
    y = .$V2,
    z = .$V3,
    type = "scatter3d",
    mode = "markers",
    color = a$rich
  ) %>% 
  layout(title = "Experiment 29")

2.2.2 Experiment 66

library(Rtsne)
set.seed(42)

tsne_df <- cdr %>% 
  # filter(!rich == "poor") %>%
  filter(time == "final") %>% 
  filter(cycle == "R0_R4") %>% 
  filter(expgroup == "66")

tsne_df %>% 
  summarise(across(everything(), ~ all(sum(is.na(.x))))) %>% 
  select(where(isTRUE))
## # A tibble: 1 x 2
## # Groups:   expgroup, cycle [1]
##   expgroup cycle
##   <chr>    <chr>
## 1 66       R0_R4
tsne_df %>% 
  filter(is.na(threshold)) %>% 
  select(rich, threshold, quantity)
## # A tibble: 0 x 6
## # Groups:   expgroup, cycle, time [0]
## # … with 6 variables: expgroup <chr>, cycle <chr>, time <chr>, rich <fct>,
## #   threshold <dbl>, quantity <int>
set.seed(42)
tsne_out <- 
  tsne_df %>% 
    ungroup() %>% 
    select(cdr3, !where(is.character)) %>% 
    select(!c(which(apply(., 2, var)==0))) %>% 
    select(!c("quantity", "cdrp", "fcp", "fcq", "rich", "threshold")) %>% 
    unique() %>% 
    mutate(cdr3 = as.factor(cdr3)) %>% 
    Rtsne(
      X = .,
      dims = 3,
      perplexity = 30,
      theta = 0.5,
      max_iter = 2E3,
      verbose = T,
      pca_center = T,
      pca_scale = T,
      normalize = T,
      partial_pca = T,
      eta = 200.0,
      exaggeration_factor = 12.0,
      num_threads = parallel::detectCores() - 2
      )
## Performing PCA
## Read the 8552 x 50 data matrix successfully!
## OpenMP is working. 6 threads.
## Using no_dims = 3, perplexity = 30.000000, and theta = 0.500000
## Computing input similarities...
## Building tree...
## Done in 8.64 seconds (sparsity = 0.016355)!
## Learning embedding...
## Iteration 50: error is 94.546345 (50 iterations in 20.90 seconds)
## Iteration 100: error is 94.053600 (50 iterations in 27.91 seconds)
## Iteration 150: error is 94.007204 (50 iterations in 13.49 seconds)
## Iteration 200: error is 94.001788 (50 iterations in 17.41 seconds)
## Iteration 250: error is 94.000470 (50 iterations in 13.39 seconds)
## Iteration 300: error is 3.809958 (50 iterations in 15.10 seconds)
## Iteration 350: error is 3.245190 (50 iterations in 9.99 seconds)
## Iteration 400: error is 2.998464 (50 iterations in 9.70 seconds)
## Iteration 450: error is 2.848536 (50 iterations in 9.44 seconds)
## Iteration 500: error is 2.747522 (50 iterations in 9.80 seconds)
## Iteration 550: error is 2.675168 (50 iterations in 9.69 seconds)
## Iteration 600: error is 2.620772 (50 iterations in 9.78 seconds)
## Iteration 650: error is 2.577785 (50 iterations in 9.89 seconds)
## Iteration 700: error is 2.543844 (50 iterations in 9.52 seconds)
## Iteration 750: error is 2.516923 (50 iterations in 9.30 seconds)
## Iteration 800: error is 2.495979 (50 iterations in 9.80 seconds)
## Iteration 850: error is 2.479166 (50 iterations in 9.81 seconds)
## Iteration 900: error is 2.465158 (50 iterations in 9.81 seconds)
## Iteration 950: error is 2.453984 (50 iterations in 10.02 seconds)
## Iteration 1000: error is 2.444859 (50 iterations in 9.56 seconds)
## Iteration 1050: error is 2.436664 (50 iterations in 9.81 seconds)
## Iteration 1100: error is 2.429402 (50 iterations in 10.33 seconds)
## Iteration 1150: error is 2.423092 (50 iterations in 10.04 seconds)
## Iteration 1200: error is 2.417693 (50 iterations in 10.41 seconds)
## Iteration 1250: error is 2.412627 (50 iterations in 10.62 seconds)
## Iteration 1300: error is 2.408001 (50 iterations in 10.42 seconds)
## Iteration 1350: error is 2.403214 (50 iterations in 11.04 seconds)
## Iteration 1400: error is 2.399291 (50 iterations in 10.63 seconds)
## Iteration 1450: error is 2.395647 (50 iterations in 10.78 seconds)
## Iteration 1500: error is 2.392276 (50 iterations in 11.01 seconds)
## Iteration 1550: error is 2.388986 (50 iterations in 10.96 seconds)
## Iteration 1600: error is 2.386141 (50 iterations in 11.30 seconds)
## Iteration 1650: error is 2.383364 (50 iterations in 11.35 seconds)
## Iteration 1700: error is 2.380737 (50 iterations in 11.50 seconds)
## Iteration 1750: error is 2.377941 (50 iterations in 11.48 seconds)
## Iteration 1800: error is 2.375171 (50 iterations in 11.11 seconds)
## Iteration 1850: error is 2.372426 (50 iterations in 11.43 seconds)
## Iteration 1900: error is 2.369839 (50 iterations in 10.95 seconds)
## Iteration 1950: error is 2.367222 (50 iterations in 10.89 seconds)
## Iteration 2000: error is 2.364811 (50 iterations in 10.61 seconds)
## Fitting performed in 460.99 seconds.
tsne_df %>% 
  ungroup() %>% 
  select(cdr3, !where(is.character)) %>% 
  select(!c(which(apply(., 2, var)==0))) %>% 
  select(!c("quantity", "cdrp", "fcp", "fcq", "threshold")) %>% 
  unique() -> a

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  ggplot() +
    geom_point(aes(V1, V2, color = a$rich))

tsne_out %>% 
  .$Y %>% 
  as_tibble() %>% 
  plot_ly(
    x = .$V1,
    y = .$V2,
    z = .$V3,
    type = "scatter3d",
    mode = "markers",
    color = a$rich
  ) %>% 
  layout(title = "Experiment 66")